Learning to smell started accepting submissions earlier this week. We noticed there are only a few people who have claimed their spot on the leaderboard till now.
We want to reach out to you all and gather early feedback for Round 2. It can help us improve your experience, we are a community driven platform after all. Please let us know however we can help!
Potential ideas:
Something you did not like in the current submission workflow
Need for more baseline implementations in starter kits
Need a helping hand while you make your first submission
Hi, I was playing with the starter kit and was able to submit entry using the random_predictor.py. However, when I tried to use fingerprint_predictor.py, the submission system gave this error message.
Submission Vocabulary contains Unknown smell words : dairy,pungent,cooling,terpenic,lactonic,alcoholic,musty,anisic,cognac,whiteflower,overripe,gourmand,ripe,cherry,geranium,pear,bread,rancid,ester,smoky,blackcurrant
One major change I did was to add environment.yml, so I could install packages using conda.
EDIT: I reverted and removed environment.yml file and the submission works. So, is there a guide on how to add environment.yml file properly?
I wanted to ask a question about new submission style. I have followed the tutorial here (https://github.com/AIcrowd/learning-to-smell-starter-kit#-making-submission), but when I look at the Issues page of my corresponding gitlab project, I get my commits labeled “failed” and this message:
“The following containers terminated prematurely. : agent. Please contact administrators, or refer to the execution logs.”
It looks like this:
May I please ask you to clarify what’s the issue here? I didn’t change any of the code. I just wanted to test and understand the new submission style. What’s funny, I was able to use this code-based commit style to make a late submission to Round 1, but it fails for Round 2. Thank you for your attention!
I have noticed a strange behaviour of the commit system which may cause the prediction setup to fail. Following the guide at https://github.com/AIcrowd/learning-to-smell-starter-kit#-making-submission, I make some changes in the files, create a new tag via git tag submission-name , and then do git push aicrowd master, git push aicrowd submission-name.
But I have noticed that the contents of my gitlab repo don’t change at all after I do these operations. So, no wonder the commits keep failing – essentially, the files in the repo don’t change either, despite me changing them locally. I’m not that proficient with git, so I may be missing something here. Could you please elaborate on this issue? Thank you for your kind attention.
